1970/71 has seen widespread use of the electron microscope in studies of void formation in materials potentially useful as fast breeder nuclear reactor fuel cladding. Whereas the gas bubble was ten years ago the major object of study now it is the void, supposed to contain no gas at all. Yet traces of rare gas, produced by neutron transmutation, evidently play a key part in void nucleation. Experimental simulations of the intense vacancy concentrations produced by high neutron fluences (<1022 ), using charged ions, show that samples of alloys preinjected with helium ions in trace quantities have immensely higher void concentrations than otherwise. The influences of alloy temperature, dispersion hardening, flux etc. are being exhaustively studied in both Europe and the USA. No acceptable growth/fluence relationship is yet available, however, partly because of the large number of parameters involved, and, in spite of the existence of satisfactory diffraction and absorption contrast theories for voids, partly because of the errors associated in measuring volume changes. Voids in fcc material are essentially octahedral, though multiple extensions and truncations can occur. Much less work has been reported on bcc material. For molybdenum the observation of a remarkably regular array of voids is described (see fig. 1). This suggests that one source of voids might be charged gas bubbles of minute size which by mutual repulsion form a crystallographic superlattice. The nucleation mechanism is the critical process to be studied here.